In recent years we have pioneered the development of fusion protein toxins. These fusion proteins combine the catalytic and transmembrane domains of native diphtheria toxin with the receptor binding specificity of polypeptide hormones and growth factors. Diphtheria toxin-based fusion protein toxins have now been constructed with alpha-melanocyte stimulating hormone, interleukin (IL)-2, IL-4, IL-6, epidermal growth factor, and athe gp120 binding domain of human CD4 as targeting ligands. In all instances, the diphtheria toxin-based fusion toxins have been shown to selectively bind to their targeted receptors, to be internalized into the cell by receptor mediated endocytosis, and to inactivate protein synthesis by the ADP-ribosylation of elongation factor 2. Since these fusion toxins are assembled at the level of the gene, we (i) have been able to optimize the cytotoxic potency of these proteins by internal in- frame deletion of portions of diphtheria toxin fragment B that are unnecessary for the delivery of th ADP-ribosyltransferase to the cytosol, (ii) have determined the position of cellular processing that is required for the release of the ADP-ribosyltransferase from the intact fusion toxin molecule for delivery into the target cell cytosol, and (iii) have recently begun to examine the structure and function of transmembrane domain regions that are required to form pores, or channels, in artificial membranes and that facilitate the delivery of the catalytic domain to the cytosol of target cells. In the current proposal, we propose to continue our studies employing site-directed and deletion mutational analysis and sub-domain substitution methods in the study of the structure function relationships of the transmembrane domain of the IL-2 receptor targeted fusion toxin DAB389-IL-2. Specifically the regions involved in pore or channel formation and the facilitated delivery of the catalytic domain across the cell membrane and into the cytosol. Since DAB389-IL-2 is extremely potent (e.g., IC50 > 5 x 10-12 M), the difference between receptor-mediated and non-specific toxicity is large (>4-logs), as a result the mutational analysis of the transmembrane domain has already proven successful.